Spectrum sensing in cognitive radio networks with known and unknown noise levels

In this study, the authors consider the problem of signal detection in cognitive radio under the cases of known and unknown noise levels. The analysis is focused on maximum eigenvalue-based method which requires the knowledge of noise variance and the ratio of maximum eigenvalue to the trace of covariance matrix. The analytical formulas are derived for threshold calculation and probability of detection for the considered methods. Performance of the considered methods is measured through simulation analysis. The performance comparison between the considered methods and other existing methods is provided. Simulations based on independent and identically distributed signals and wireless microphone signals are presented to verify the various sensing methods.